WGU D570 STUDY GUIDE
LATEST MOCK PRACTICE SET
139 Questions with Answers and Detailed Rationales
100 PERCENT GUARANTEED PASS
INSTANT DOWNLOAD ANSWERS INCLUDED
IMPORTANCE OF THIS DOCUMENT
This comprehensive examination preparation guide has been meticulously developed to help you succeed in the
WGU D570 STUDY GUIDE. It contains 139 carefully selected questions that reflect the most current exam
content and testing strategies. Each question is accompanied by a correct answer and a detailed rationale that
explains the underlying pathophysiology, pharmacology, or clinical reasoning.
Self-Assessment – Test your knowledge and Exam Preparation – Familiarize yourself with the
identify areas requiring further question format and content
study areas
Concept Reinforcement – Deepen your Confidence Building – Develop test-taking
understanding through strategies and reduce
evidence-based exam anxiety
rationales
Time Management – Practice answering
questions under simulated
exam conditions
Review Summary 139 Questions
Foundations - Application - WGU D570 Study Guide DATA Analytics AND Machine Learning Graduate
All answers with rationales
,Table of Contents
Section A - DATA Management AND Section B - DATA Quality AND
Governance Integration
Questions 1 to 35 Questions 36 to 70
Section C - DATA Warehousing AND Section D - DATA Analytics AND
Business Intelligence Visualization
Questions 71 to 105 Questions 106 to 139
,Section A - DATA Management AND Governance
Q1.
In a high-dimensional dataset with 5000 features and 200 samples, which regularization
technique minimizes the risk of overfitting while simultaneously performing feature
selection?
A. Ridge regression with alpha=0.1 B. Lasso regression with alpha=0.01
C. Elastic Net with L1 ratio=0.5 D. Principal Component Regression with 50
components
Correct: C - Elastic Net with L1 ratio=0.5
Rationale:Elastic Net combines L1 and L2 penalties, making it effective for feature selection
and handling groups of correlated features. Lasso (B) can select at most n features, which is
suboptimal here. Ridge (A) does not perform feature selection. PCR (D) does not directly
select features.
Q2.
A data scientist trains a random forest model on a dataset with class imbalance (1:100
ratio). The model achieves 99% accuracy but only identifies 5% of the minority class.
Which metric should be prioritized to evaluate model performance?
A. F1 score B. Area under the ROC curve (AUC-ROC)
C. Precision-recall AUC (AUC-PR) D. Cohen's kappa
Correct: C - Precision-recall AUC (AUC-PR)
Rationale:AUC-PR is more informative than AUC-ROC for imbalanced datasets because it
focuses on the minority class. F1 score (A) is a threshold-dependent metric and may not
capture overall performance. AUC-ROC (B) can be overly optimistic with high imbalance.
Cohen's kappa (D) measures agreement but is less standard for imbalance.
Q3.
In a deep learning model for image classification, the training loss decreases steadily but
the validation loss starts increasing after epoch 10. Which combination of actions is most
appropriate to address this?
A. Increase learning rate and add dropout B. Reduce model complexity and implement
layers early stopping
C. Use data augmentation and increase D. Add more convolutional layers and
batch size reduce regularization
Page 3
, Section A - DATA Management AND Governance
Correct: B - Reduce model complexity and implement early stopping
Rationale:Increasing validation loss indicates overfitting. Reducing model complexity (e.g.,
fewer layers or units) and early stopping directly combat overfitting. Increasing learning rate
(A) or adding more layers (D) may worsen overfitting. Data augmentation (C) helps but is less
direct than reducing complexity and early stopping.
Q4.
Given a time series with strong seasonality and trend, which approach would yield the
most accurate forecasts when the seasonal pattern changes over time?
A. Holt-Winters exponential smoothing with B. SARIMA with fixed seasonal period
additive seasonality
C. Prophet with changepoint detection D. ARIMA with differencing
Correct: C - Prophet with changepoint detection
Rationale:Prophet is designed to handle changepoints in trend and seasonality, making it
robust to evolving seasonal patterns. Holt-Winters (A) and SARIMA (B) assume fixed
seasonal patterns. ARIMA (D) does not model seasonality without differencing.
Q5.
A logistic regression model yields a coefficient of 1.5 for a binary feature X. What is the
odds ratio for Y=1 when X=1 compared to X=0, assuming all other features are held
constant?
A. 1.5 B. e^1.5
C. ln(1.5) D. 1/(1+e^-1.5)
Correct: B - e^1.5
Rationale:In logistic regression, the exponentiated coefficient gives the odds ratio. For a
coefficient , the odds ratio is e^. Thus, e^1.5 is correct. Option A is the log-odds, C is the
inverse, and D is the probability when X=1 with zero intercept and other features zero.
Q6.
In a clustering analysis using k-means, the elbow method suggests k=3 but the silhouette
score is highest at k=5. Which evaluation metric is more reliable for determining the
optimal number of clusters when clusters have non-spherical shapes?
A. Elbow method using inertia B. Silhouette score
C. Davies-Bouldin index D. Gap statistic
Correct: D - Gap statistic
Page 4
LATEST MOCK PRACTICE SET
139 Questions with Answers and Detailed Rationales
100 PERCENT GUARANTEED PASS
INSTANT DOWNLOAD ANSWERS INCLUDED
IMPORTANCE OF THIS DOCUMENT
This comprehensive examination preparation guide has been meticulously developed to help you succeed in the
WGU D570 STUDY GUIDE. It contains 139 carefully selected questions that reflect the most current exam
content and testing strategies. Each question is accompanied by a correct answer and a detailed rationale that
explains the underlying pathophysiology, pharmacology, or clinical reasoning.
Self-Assessment – Test your knowledge and Exam Preparation – Familiarize yourself with the
identify areas requiring further question format and content
study areas
Concept Reinforcement – Deepen your Confidence Building – Develop test-taking
understanding through strategies and reduce
evidence-based exam anxiety
rationales
Time Management – Practice answering
questions under simulated
exam conditions
Review Summary 139 Questions
Foundations - Application - WGU D570 Study Guide DATA Analytics AND Machine Learning Graduate
All answers with rationales
,Table of Contents
Section A - DATA Management AND Section B - DATA Quality AND
Governance Integration
Questions 1 to 35 Questions 36 to 70
Section C - DATA Warehousing AND Section D - DATA Analytics AND
Business Intelligence Visualization
Questions 71 to 105 Questions 106 to 139
,Section A - DATA Management AND Governance
Q1.
In a high-dimensional dataset with 5000 features and 200 samples, which regularization
technique minimizes the risk of overfitting while simultaneously performing feature
selection?
A. Ridge regression with alpha=0.1 B. Lasso regression with alpha=0.01
C. Elastic Net with L1 ratio=0.5 D. Principal Component Regression with 50
components
Correct: C - Elastic Net with L1 ratio=0.5
Rationale:Elastic Net combines L1 and L2 penalties, making it effective for feature selection
and handling groups of correlated features. Lasso (B) can select at most n features, which is
suboptimal here. Ridge (A) does not perform feature selection. PCR (D) does not directly
select features.
Q2.
A data scientist trains a random forest model on a dataset with class imbalance (1:100
ratio). The model achieves 99% accuracy but only identifies 5% of the minority class.
Which metric should be prioritized to evaluate model performance?
A. F1 score B. Area under the ROC curve (AUC-ROC)
C. Precision-recall AUC (AUC-PR) D. Cohen's kappa
Correct: C - Precision-recall AUC (AUC-PR)
Rationale:AUC-PR is more informative than AUC-ROC for imbalanced datasets because it
focuses on the minority class. F1 score (A) is a threshold-dependent metric and may not
capture overall performance. AUC-ROC (B) can be overly optimistic with high imbalance.
Cohen's kappa (D) measures agreement but is less standard for imbalance.
Q3.
In a deep learning model for image classification, the training loss decreases steadily but
the validation loss starts increasing after epoch 10. Which combination of actions is most
appropriate to address this?
A. Increase learning rate and add dropout B. Reduce model complexity and implement
layers early stopping
C. Use data augmentation and increase D. Add more convolutional layers and
batch size reduce regularization
Page 3
, Section A - DATA Management AND Governance
Correct: B - Reduce model complexity and implement early stopping
Rationale:Increasing validation loss indicates overfitting. Reducing model complexity (e.g.,
fewer layers or units) and early stopping directly combat overfitting. Increasing learning rate
(A) or adding more layers (D) may worsen overfitting. Data augmentation (C) helps but is less
direct than reducing complexity and early stopping.
Q4.
Given a time series with strong seasonality and trend, which approach would yield the
most accurate forecasts when the seasonal pattern changes over time?
A. Holt-Winters exponential smoothing with B. SARIMA with fixed seasonal period
additive seasonality
C. Prophet with changepoint detection D. ARIMA with differencing
Correct: C - Prophet with changepoint detection
Rationale:Prophet is designed to handle changepoints in trend and seasonality, making it
robust to evolving seasonal patterns. Holt-Winters (A) and SARIMA (B) assume fixed
seasonal patterns. ARIMA (D) does not model seasonality without differencing.
Q5.
A logistic regression model yields a coefficient of 1.5 for a binary feature X. What is the
odds ratio for Y=1 when X=1 compared to X=0, assuming all other features are held
constant?
A. 1.5 B. e^1.5
C. ln(1.5) D. 1/(1+e^-1.5)
Correct: B - e^1.5
Rationale:In logistic regression, the exponentiated coefficient gives the odds ratio. For a
coefficient , the odds ratio is e^. Thus, e^1.5 is correct. Option A is the log-odds, C is the
inverse, and D is the probability when X=1 with zero intercept and other features zero.
Q6.
In a clustering analysis using k-means, the elbow method suggests k=3 but the silhouette
score is highest at k=5. Which evaluation metric is more reliable for determining the
optimal number of clusters when clusters have non-spherical shapes?
A. Elbow method using inertia B. Silhouette score
C. Davies-Bouldin index D. Gap statistic
Correct: D - Gap statistic
Page 4
